Research On Active Steering-by-wire Control Strategy Of Driverless Mining Trucks

With the development of science and technology,self-driving technology is becoming the research focus in intelligent transportation field.In recent years,self-driving in mining areas has attracted public’s attention,which is one of the scenarios most likely to achieve commercial application.In the meanwhile,there are many difficulties in practical application.This thesis focuses on the steering control of driverless mining trucks,which aims to realize accurate tracking of reference path.Considering the effect of time delay of truck steering actuator,the control algorithms were improved to meet the requirements of practical applications.In view of structural characteristics of trucks,the truck kinetic and dynamics models were built.The steering time-delay system was built to to simulate the time delay of steering actuator.Based on preview control,full state feedback control and model predictive control theory,a variety of controllers have been designed.Simulation verification was carried out on Truck Sim/Simulink co-simulation platform.Finally,the field test was carried out to verify the algorithm.Firstly,system models were built.In view of the dual-wheel combination structure of trucks,the truck dynamics model in plane was built,which was then simplified in the absence of load transfer.The vehicle error control equations were derived.To simulate the time delay of the actuator,steering time-delay system was bulit.A second order system was adopted to simulate the execution process of the truck’s steering system.Based on preview control theory,the preview PID control algorithm was described.The improved adaptive preview control algorithm was designed.Based on the truck dynamics model,the linear quadratic regulator(LQR)controller and model predictive controller(MPC)were designed for steering control.Against the steering time-delay system,the preview LQR controller,predictive LQR controller and cascade LQR-PID controller were designed to improve the LQR controller.Combining model reference adaptive control(MRAC),the MPC-MRAC controller was designed.The Truck Sim/Simulink co-simulation platform was built.The HIL test platform was described.Combining test datas,the truck dynamics model and steering time-delay system were verified.The results show that the system models have good accuracy.Under the dual-shift condition,the designed controllers were verified by simulation.The simulation results show that improved controllers could significantly reduce the max and mean tracking errors and improved vehicle stability in the steering time-delay system at higher vehicle speeds,which have better performance.Besides,under the actual road path,the improved adaptive preview control algorithm,LQR-PID controller and MPC-MRAC controller were simulated and compared.The results show that MPC-MRAC controller has lowest errors in general,which means better performance.Finally,the overall architecture of driverless mining trucks in this thesis,route-collected planning and coordinate conversion methods were decribled.Under curve and straight road conditions in mining area,the designed improved adaptive preview control algorithm was applied to the field test.The test results verified the feasibility of the improved adaptive preview control algorithm and influence of preview points numbers on tracking performance.The test tracking errors show that the tracking errors of improved adaptive preview control algorithm can meet the requirement of practical application at lower speeds.The steering control strategies in this thesis considered the problems of the algorithm in practical application and were verified mainly by simulation.Some algorithms were applied to the field test.The research achievements have certain theoretical and practical application value.